Loading summary

Transcript64 lines

[00:00]
A
Foreign.
[00:09]
B
Welcome to the Corporate Director Podcast where we discuss the experiences and ideas behind what's working in corporate board governance in our digital tech fueled world. Here you'll discover new insights from corporate leaders and governance researchers with compelling stories about corporate governance strategy, board culture, risk management, digital transformation and more.
[00:33]
C
Hi everybody and welcome back to the Corporate Director Podcast, the voice of modern governance. My name is Dottie Schindlinger, Executive director of the Diligent Institute, and I'm joined once again by my two amazing co hosts, Megan Day, strategy leader here at Diligent, and Kira Ciccarelli, who's the head of research at Diligent Institute. Megan, Kira, how are you both doing today? Very well, Dottie.
[00:55]
D
Pretty good.
[00:56]
C
So hopefully everybody had a nice weekend. I know here in the northeast of the U.S. we've just come through lots of bad air quality from the really awful, awful fires up in Ontario and then we had major storms again. Boy, this has been, this has been a summer for the record books, hasn't it? It's been crazy here.
[01:13]
D
And by the way, don't buy the produce, I guess for a little while.
[01:16]
C
I know on Saturday we had this idea of like going to a local salad bar and then three seconds later went, maybe not the time to be eating fresh lettuce. Maybe in a couple weeks. Maybe in a couple weeks. But listen, Kira, I want to give you the floor because you flagged us on an article that you saw in today's CEO Daily about Jeremy Allaire's treatise on the agentic economy. Do you want to kind of talk about that a little bit?
[01:40]
D
Yeah. So this absolutely caught my eye this morning when I was going through my newsletters and my inbox this morning. And essentially, Jeremy Allaire's tech company CEO, he and his team have put together a sort of treatise on the agentic economy. And basically the point that he is making here is that as AI agents become more dominant in workflows and just in the economy and business world in general, they're going to have to start operating with sort of blockchain and things like that, as opposed to traditional means of changing money back and forth. And it was really, really fascinating.
[02:24]
C
Yeah, I have to be honest, there's lots of different versions you can read. So if you go to the website and you check out this treatise, there's the 62nd thesis, there's the Tell it to Me like I'm a 12 year old version, there's the long read, there's the podcast, there's a video. I went to the Tell it to Me, like I'm a 12 year old version even that I found really challenging. But the thing that I was struck with. So he's talking about agents managing agents and you know, dividing jobs into discrete tasks that they can assign to each other. And the way that it's written made me think, so is this the end of corporations as we know it, part one and part two? Where in this treatise does he cover what happens to the rest of us? Because he talks about how, you know, agentic AI can't replace hands on work. So people who, you know are caregivers or, you know, sanitation workers or anybody who's working with their hands, their jobs are probably safe for now because robotics is still a long way away from replacing them. And all the people at the top are okay because they, you know, you still have to have a human being taking accountability, responsibility. But the rest of us, question mark, did he cover it anywhere in the street?
[03:41]
D
So I like you, I started with the 62nd version and then I was thinking to myself, none of these words are making any sense to me at all. So I quickly went over to the. Explain it like I'm 12. And yeah, I think you're right. I mean, he makes the point in this five minute version that's kind of covering some of the major tenets of the treatise, that the huge risk here is exactly what you're saying. If everything is handed off to agents and you only need one person overseeing a mass of agents, then it is very likely that power will become very concentrated in the hands of very few people.
[04:27]
C
So that doesn't sound like a great idea to me. I, you know, it's, it's really interesting as I read through it, as I thought about it some more, I mean, in fairness, Jeremy Allaire, CEO of a company that is in the blockchain space. So maybe not super surprising that he's anticipating that blockchain will be part of the ecosystem that makes this all possible and happen. But there were certain phrases that really kind of set me back, like this will happen and it'll seem like it was almost by accident. There was a phrase in there that said that, and I thought, you know, that is really important for us to pause on and say that will be true if we don't do anything right now. Megan, I don't know what your take on all this is, but yeah, what
[05:13]
E
I struggle with, and this has come up in some of our other conversations is how soon is this reality going to be here? And, and I oscillate between, oh my God, the End is near to. All right, we have some time here. This is, you know, 10, 15, 20 years out before we live in this society. I do encourage folks, if you haven't read it, earlier this year in the New Yorker, there was a great piece called what is Claude Anthropic? Doesn't know either. And it's worth reading on a couple of fronts. One, it sends you into the mind, if you will, of this agent economy and some of the things that we might grapple with in the future. But like the manifesto, the treatise, it does pose a million different questions about where we're headed and how quickly we will get there.
[06:11]
C
I mean, one of the things I think about, and this is probably a really good segue for us to listen to our interview for today, but one of the things I think about is, you know, we're seeing signs that this is already starting. And so to your question, Megan, about how soon is it before we're in this new world? Well, we will get there. If we don't put guardrails in place right now. Like, if we don't, if we don't, like, have some conversations at the right leadership level to talk about how do we keep this from being an existential threat, then it will be an existential threat within a generation. Like, within a generation. If what this treatise lays out comes to pass and nothing else changes, that's the most important thing. If nothing else changes within a generation, you have absolutely no pipeline for future leadership. Within a generation, you have absolutely no one who is prepared and qualified to be those people at the top running the agent economy.
[07:03]
D
Right.
[07:04]
C
Truthfully. Because how do we, how do we prepare people today? They come up through the ranks, they learn entry level jobs, then they get more responsibility, then they manage some people, then they get more responsibility, and over the course of years, they accrete enough experience and an education that they are now ready to lead teams, departments, companies, that whole pathway. If we were to just allow the agentic economy to happen the way it's written about in the treatise completely breaks down and whole companies, whole types of jobs, whole sectors of the economy all become agentic with robots governing other robots. What are the people doing? I'm seeing you both nod your heads and saying absolutely nothing, which is terrifying.
[07:49]
E
To even remind me of the other manifesto, if you will, from earlier this year. And I'm blanking on who wrote it, but we talked about it in the podcast.
[07:57]
A
That's it.
[07:58]
E
I mean, it's terrifying as it is exciting.
[08:03]
C
So that's probably a very good segue to A Glimmer of Hope, which is the event that Kira and I got a chance to go to in June in Brussels. We were invited to come and speak at the Digital Trust Summit that was organized by Dominique Shelton Leipzig. And so today we sat down with Dominique, we asked her lots of questions about the event, about why it's important, about why every director should think about trying to go next year, every CEO should think about trying to go next year. And this treatise I think just adds extra fuel to that fire for me that we really all, we will all do ourselves a disservice if we don't get this right. And pretty much right now we need to get the right guardrails in place. And the great thing about this event was it was all about guardrails and innovation. It wasn't about how guardrails slowed down innovation, it was about how guardrails speed innovation. And I think that's really cool. So why don't we give a listen to the interview and we can come back and maybe the two of you will have more to say. Joining us on the Corporate Director podcast today is Dominique Shelton Leipzig, founder and CEO of Global Data Innovation. Dominique, thank you so much for joining us again on the show.
[09:17]
A
Dottie, thank you so much for having me. It's great to be back.
[09:20]
C
Well, let's just get started off here by having you introduce a bit more about yourself to our listeners than your current title. You've had quite an amazing career, so could you share a little bit about your career background?
[09:32]
A
Yeah. So you know, I started my company, Global Data Innovation almost three years ago now because I could see that we were getting with AI and new technologies kind of in the same pickle we had gotten into with cyber. In other words, there is new technology out there, confusion at the board level and CEO level about what to, what to do with this new technology, how to govern it. And then what I noticed was a tendency to over delegate to anyone else that would take this on to figure it out. And what that leaves is a huge bucket of risk. So from those days in Cyber 25 years ago, I kind of vowed to myself I would never let that happen again. Because of course, cyber has turned into an issue that has cost our global economy about 12 and a half trillion per year. And in the past 10 years, over 60 trillion has gone in cyber related fines and expenses and ransomware payments, et cetera. And when I saw those trends starting to happen with AI about three years ago, I realized that we needed to do something we didn't do with cyber, which is leverage best practices solutions and really our study of AI failures, bringing those learnings to our leaders ahead of time so they could map solutions to avoid those problems and be successful with their technology.
[11:05]
C
Well, I know that the work that you're doing is really having a big impact because myself and Kira Ciccarelli, my colleague, had the great privilege to join you in Brussels just a few weeks ago at the Digital Trust Summit. And I think this is the fourth year, is that correct? Maybe the fourth or the fifth year?
[11:23]
A
The fourth, yeah.
[11:25]
C
So we were so excited to be there. Could you share with our audience a little bit more about the goals of the summit, kind of what led you to start this event in the first place and what were some of the biggest themes or wake up calls that were shared at this year's event?
[11:39]
A
Yeah, so Donny, I started the Digital Trust summit in 2023, very close, about a year before I started my company. And I started it for the very same reason I started my company. I could see, see that our CEO and board community were a little bit deers in headlights with the new AI technology and through no fault of their own. The technology is moving so fast, the trends are moving so quickly and there were so many headlines even back then. It was clear to me that we needed to create a space where our CEOs boards could get together with regulators, legislators and large language model experts, data scientists and academia, and put our heads together about solutions not hyped. So that mission that started in 2023 continues and remains as you know, the Dottie, because you do so much of this work at the Diligent Institute, we are literally our, our CEO and board members are inundated with all kinds of signals, headlines, noise. It's very hard to make, make sense of it all and to figure out what should boards be doing right now to govern AI. So the. The goal of this conference was really to provide in clarity and a roadmap for solutions for excellent exercise of fiduciary duty over AI. And I believe we accomplished that. So I'm very proud that you were there and you were a big part of that. Dottie, in terms of your panel on laying out what boards in the EU and US need to know about trends in order to arrive at great governance.
[13:17]
C
Well, speaking of boards in the EU and the us we are just days away from a big milestone in the EU AI Act. I know August 2nd represent a really important day, so could you share a little bit about what directors need to know most about the EU AI act and what's happening on on August 2nd that they need to be aware of?
[13:38]
A
Yes. So for directors that are directors of companies that either do business in the eu, have vendors in the eu, or see the EU as a growth market, which is a big swath of our US companies, if you're on a board of a company like that, the EU AI act should be on your radar screen on August 2nd. First of all, the EU AI act is a sweeping law. There's over 1,000 KPIs in it, things that are very hefty, about 700 pages or so of law in there, a lot going on. But what we've done at Global Data Innovation is taken that law and looked at what aspects of this law are most germane to avoid the tens of thousands of incidents that we've been studying. As I mentioned, there are so many incidents that have occurred with AI in the four years since ChatGPT hit the market. And from those learnings, we've seen that out of those thousand KPIs in the EU AI act, there are only about five that really are moving the needle and causing about 85% of the incidents that have been documented. The tens of thousands of incidents out there. So what we wanted to do with what we want directors and CEOs to focus on at the level that they're in, obviously their legal counsel will make sure that they're buttoned up and all the legal compliance. But what a CEO and a board member needs to know, what are the things that are going to, that deserve my attention so that I can avoid or help our management team think through how to avoid losses that have happened to maybe 95% of the companies deploying AI, so that we do not repeat those incidents in our company? So we want incidents to stop where they are, not repeat themselves over and over again in companies in various ways, like what we have with cyber.
[15:42]
C
So, you know, a couple of things that you just touched on there are really resonating for me because I think what always happens when we have a new technology come to play, the development of that technology and the adoption of the technology almost always outstrips the governance of the technology. And I think we're absolutely seeing that with AI, that AI is moving a lot faster than regulation is able to keep up, and really, even organizational governance is keeping up. So what are some of the questions that boards should be asking management right now to really get a sense of where AI is being used across their businesses, including shadow or embedded AI? What, what should they be asking about yeah.
[16:23]
A
You know, first, every board member should be asking for management to make sure that AI is on the agenda in a specific way. To, to your point, do what AI use cases do we have going on in the business? That is a great question to get a report on. And then similarly, once that report is delivered, having the resilience or pushback to ask, what do these AI use cases that are in flight? What do they have to do with our strategy in the company in terms of revenue, operations and overhead cost reductions? What impact will this AI have on our business if it is successful? So getting aligned on those preliminary questions really helps the board get a sense of are we doing AI that's really going to move the needle if it's, if it's successful, or are we doing science projects in the company that even if we have home runs with it, we're not going to be able to demonstrate to an investor or to our stakeholders, our business partners and employees and customers that this is added value.
[17:29]
C
So, Dominique, one of the things I was eager to talk to you about is the trust framework that you developed. And trust T rust is an acronym that stands for triage.
[17:39]
D
Right.
[17:40]
C
Data, uninterrupted monitoring, supervision and technical documentation. And first of all, I love that framework so much because it's so easy to understand and get your head around. But I wanted to ask you more specifically about the genesis of that framework and how that maps to the EU AI Act. And also just where do you see companies falling short on that framework today?
[18:06]
A
Yeah, so the genesis of the framework was really ground up. I looked at AI incidents that were occurring and then compared them with where I saw commonalities in the legal frameworks in 100 countries, six continents. A lot of those frameworks are on my company's website@globaldatainnovation.com there's a frameworks tab and you can see all the countries that have legal frameworks. And what I noticed is there were certain themes that kept coming up in all of the legislation. I also then we took the four years now since almost four years of data we have since ChatGPT4 was first announced. And there are tens of thousands of incidents. So what we did is we mapped where we saw commonality in the law with what would have prevented those tens of thousands of incidents from occurring had those best practices, some of the laws aren't in. In effect yet. Had those best practices been enacted, what would have prevented those losses? And it came down to these five areas. You know, when I think of the EU AI act, which has over a thousand KPIs or NIST, which many companies are mapping to just The NIST Gen AI framework alone has over 260 KPIs. That's so much for a board or a CEO or even your general counsel to find focus on all at once. So I wanted to find a clean way to prioritize what is really important. And what we learned is, is that about 85% of the incidents are caused by the lack of something in the T area, the R area, the U area, the S area, the T area that you outlined, Dottie. So that's why we focus on it, because we figure that if a board or a CEO focuses their attention on innovating to anticipate and avoid those issues, then they' to be on the road to success with their AI.
[20:07]
C
So I love that so much. You know, of course if you're on a board, you're going to want to ask what kind of framework is in place around governance of AI. But then are there a couple of the letters that board members should be asking about at every meeting? I mean, are there specific things they should be getting reports on at every meeting?
[20:25]
A
Yeah, I think the two most important and where we've seen the biggest losses are in the area of R, the right data to train. And that's really about data accuracy, making sure that there's a process for ensuring that accurate data is used for the AI. And then second in the right data has to do with IP rights, privacy rights and business rights to train. We're seeing billions of dollars going into settlements on for copyright and other IP related claims. And so it's an area of focus that we think boards should pay attention to. Another area for that the board should pay attention to here is, is and I call it like the fail safe is the untrust, uninterrupted testing, monitoring and auditing.
[21:12]
C
And I would imagine part of the reason why boards should be asking about those two areas is that especially in terms of the uninterrupted monitoring, that that might sound like something that is costly and so there might be pushback on it. But what would, what would you counsel a company that tells you such a thing?
[21:31]
A
Well, the good news is times have changed. AI is here and uninterrupted testing, monitoring and auditing. I've seen certain guardrails, you know, say there's monitoring against a known accuracy standard that the company has for say if it's a genomic company for certain types of R and D for disease, or if it's a retailer, certain standards they have for customer service. The good news is that with AI, these standards can be coded into the AI itself without using generative AI. They can be used using deterministic or what I mean by deterministic, just hard coded rules that don't need to shift with the Internet or the waves of the whole wide Internet. These are your company's standards with respect to accuracy. So they can live in uninterrupted testing, monitoring and auditing in a hard coded way. They can live either in what something called an observability model, or they can live in the AI itself as a monitoring vehicle. And we've seen that done as low as five figures, six figures. So let me put this in context. If you're working for, you're on the board of a big bank or a healthcare company, they are putting in, say it's a small mid cap healthcare company, they're putting in millions of dollars associated with AI development and some of our bigger banks anywhere between, you know, the proxies reveal and their annual reports, anywhere between, you know, 15 billion, 19 billion in digital transformation and AI development. In that context, a six figure guardrail is literally a drop in the bucket. And when I think of the CEOs just this year that have had to step down because of AI incidents and the shareholder derivative actions that are naming board members individually for failing to exercise oversight and the months that those directors will be sitting in deposition and looking through their emails and sitting with counsel, it is so worth it to take care of these issues and ask the question as a board member now rather than be defending your governance and your oversight later.
[23:50]
C
Such good advice. Well, one of the other aspects of AI and governance is the use of AI in board work itself, which of course is something that we spoke about at the Digital Trust Summit. Diligent Institute's been doing a lot of research in this area and one of the things that we have found is that while directors are using AI pretty, pretty frequently to help them with board work, very few companies have AI usage policies in place that governor usage of AI specifically in the boardroom. So if you were advising a board that you know, has been experimenting a little bit with AI, but still doesn't have a very mature governance framework, what would be some of the first things that you'd recommend directors do in the Next, let's say 60 to 90 days to really get ahead of the risk?
[24:34]
A
Well first, within the first 30 days, I would highly recommend that the board get to the Diligent portal and pull down the sample template AI Board policy that you've made available, Dottie, which is incredible, but it's absolutely important for the board to have their own AI governance policy. Why? Because in the interstitial area of boards using AI and the lack of guidance on what it should be used for and should not be used for, risk stands. You know, think about this. We already know of several public companies where this has happened, where they've board members have used AI commercially available AI, not enterprise licenses, and competitors have been able to derive intelligence about those boards and about the companies that those board members are on by virtue of the prompts, even if they don't even put in confidential information by virtue of the prompts that they use. So this is so important for boards to have that guidance and to focus on making sure that there's a policy in place. Second step is training on that policy, so everybody knows why it's there. Dottie, you and I have done a number of trainings together and I, I. One of the things I like doing with you so much is the fact that both of us bring up a lot of stories. I find it so much easier for a board member to remember the importance of why it's important to use an approved portal that has security controls like what you've got going on at Diligent versus the general commercial tools that are available and they're not necessarily confidential. And the providers themselves say, please don't put anything in here or use prompts that you would be concerned about being the subject of a data breach, because many of the providers have suffered data breaches. That's the reason they tell you that in their terms. So I love to do that storytelling. So our board members have the, not just the, the policy itself, but the why it's important and what it's there to protect.
[26:47]
C
Such, such good advice. Well, listen, before we wrap up our conversation here, there are three questions that we'd like to ask every guest that joins us on the show. And I know you've answered these before, but we're going to ask them again. So the first question is, what do you think will be the biggest difference between boardrooms today and 10 years from now?
[27:05]
A
Well, I think you're going to see the boards that are around 10 years from now are going to be the ones that adopt these proactive practices that I'm talking about here. There's about to be a huge schism that lays out in the next, I would say 12 months to 24 months between those boards that proactively lean into the technology and also equally Important lean into the governance so that they're turning out AI products that have a level of, that match the brand, that have a level of excellence, accuracy and really reflect what's best about your brand. There'll be the others that are not going to make it because they're going to turn out products and AI enabled material that's inaccurate, that causes brand tarnishment and harm to their customers, employees and business partners. And they're not going to be sitting with us ten years from now. So I'm excited that we're going to have a really sophisticated group of companies and boards that have learned how to manage this technology and many more new things. We're just at the beginning of Quantum and we're at the nascents of what AI will be doing in agents and agentic. AI and robotics saw so many prototypes at the UN and Geneva of all the new technology that's coming our way. But the boards that become flexible in this reflex of being continuous learners, asking questions, not deferring and putting our heads in the sand and over delegating to the technical folks, but jumping in with strategy and making sure that you understand as a board member what you're overseeing. Actually ask for the types of briefings that will help you understand it. That's the. Those are the boards of tomorrow and I'm excited about that.
[28:51]
C
Well, speaking of briefings, what was the last thing that you read or watched or listened to that made you think about governance in a new light?
[28:58]
A
There are two things. One is the Pope's encyclical letter Humanitas Magnificus, which came out in June. And I just. It's about 100 pages long. It's a long read and when you read it, it's. You can see some of the difference between AI writing and a real person's writing. Because it's clear that somebody working with the Pope and the Pope himself spent a lot of time in putting together a great document that I think eventually history will show as a roadmap. It's not really a religious document. It's about planning and thinking ahead of the society that we want with AI and how to get us from point A to point B. It talks about the dignity of work and also the importance of accuracy and truth. These are all things that are going to be essential for our future. And I'm excited that there's roadmap there in that document. So I found that that was very synergistic with the work that we do at Global Data Innovation and also an extension of which is our Digital Trust Summit where we act out of simulations and work on forward planning and thoughtful pre planning around AI. The other book I came back to is Haruki Murakami's IQ84. I highly recommend it. It's really about technology, lives, science fiction. It also has kind of a Proustian way about it. It's another long read. That one's about 1200 pages. But I read it about 10 years ago and I came back to it and I've been rereading it this summer and it's just been a joy.
[30:38]
C
Those are great, great suggestions. I think you get the prize for being the first guest to ever recommend a papal encyclical. So thank you for that. That's really fantastic. And Dominique, before we let you go, what is your current passion project?
[30:53]
A
Yeah, you know, right now I'm really passionate about making sure that we build also the next generation of governance experts. So I started a Career Ascend program as part of Global Data Innovation. We just came off our first cohort of 11amazing governance professionals and really refined many of the things that we've been talking about here. How to use their voice, how to see something, say something, and if there's something that they notice at the ground level, to be sure to raise those things to the CEO and board so that everybody's working off the same sheet of music and we're getting to great governance, not waiting for it to be turned into something like what happened with cyber, where we kept things in silos and it ended up costing our global economy trillion. So I'm excited about Career Ascend.
[31:47]
C
Well, Dominique, I'm excited that I got a chance to speak with you again. Thank you so much for coming back and joining us on the show.
[31:52]
A
Thank you for having me. Dottie, it's been an absolute pleasure and congratulations on all the work you're doing at Diligent.
[31:58]
C
We've been joined once again by Dominique Shelton Leipzig, founder and CEO of Global Data Innovation. Dominique, thank you so much for joining us.
[32:06]
E
I was going to say stop talking. Yes, this is all good, Kira. I feel like you got it.
[32:14]
D
Great interview, Dottie. It's always such a pleasure to get to talk to Dominique and just kind of mine her brain. And I think one thing that really stuck out to me from this interview in particular is that she comes back often to this idea of we have a chance now to get it right and to learn from previous instances where new regulations or new technologies came to pass and maybe we weren't as prepared and thinking about this. Treatise on the agentic economy that we've been discussing. My first thought reading it was when you're talking about agents, AI agents operating within a capitalistic society where it's money changing hands and you have agents doing that work. I just imagine every cfo, every legal leader, just like, I can almost hear them screaming as they read it, like, from where I am. And I think that's just such a great example of where something like the trust framework, which Dominique talks about could be so, so important. And the idea of knowing the art of the possible and what might be coming and be able to set up those guardrails in advance. We're at the point now where we have the chance to do that.
[33:27]
C
These are times when this is totally going to date me, but these are times when I hear Jeff Goldblum from Jurassic park in my head saying, you were so focused on whether or not you could, you didn't pause to ask if you should kind of feel like that applies. Megan, what was your big takeaway from this interview?
[33:46]
E
Well, I will. I mean, I come back to what we talked a little bit before about governance not being anti innovation, that it can, you know, if you have weak governance, it creates paralysis and you can't scale AI enterprise wide. But she gives some really practical advice too, for boards about how to think about this, which I think again, when we talk about all of this sort of nebulous stuff, you really need to come back to, well, what can I do right now? And that's about asking at every meeting whether the company has uninterrupted testing, monitoring and auditing of these AI outputs. I mean, I think that is just, again, what we need to keep coming back to.
[34:33]
C
And I really loved her point that, you know, you might get an answer of, well, that's too expensive and cumbersome and we can't possibly do it. And she, she brought up the case of, you know, I've seen it happen for as little as five figures or six figures. She's like, you know, if you're making these multi billion dollar bets, that's like, that's couch cushion money, you know, to, to invest in having something that will keep you from an absolute existential threat. I think the other thing I always appreciate about Dominique's take on how we can get this right now, you know, she lived through her career spans across the entirety of the cybersecurity world and she saw how there was a moment in the 90s when we could have gotten ahead of it and had we put some frameworks in place at that point in time, we wouldn't have CyberCrime as a $12 trillion a year industry right now. And I think nobody wants a repeat of that. Like literally no one wants something like that to happen with AI. Nobody. That's not good for anyone except maybe the criminals. It's not good for anyone else. And it has caused nothing but headaches for companies, for consumers, for employees, for everyone. And I think that's such a good analogy. And I just keep holding onto that because I think that's what's in it for us.
[35:45]
D
Right.
[35:46]
C
You know, to your point, I don't think good governance slows down innovation. I personally think good governance is good governance when it fuels innovation, when it allows innovation to move on smoother Rails. Because you have rails. Like if you don't have Rails, it can just spiral out of control and pretty soon you end up with a company that is ending because it's done something really terrible without your intention, and now it's terrible. And now you have to clean it up versus good rails allow you to go smooth and fast in a direction.
[36:18]
E
And granted, I am not in a public company boardroom at the moment, but I have to think a lot of the conversations are focused on perhaps you need to be using AI more. Add it to this use case. Add it to that use case around the business versus coming at it with a skeptics lens of how does this connect to strategy, how does this connect to operations? How is this creating value for the organization outside of maybe cutting heads? So lots to think about.
[36:50]
D
I hear that all the time. And I think we see a lot of organizations move in the wrong order. Like they start with, we have this tool now, so where can you force fit this in? Where can you plug this in? Instead of starting with what are we doing? That is tedious and takes a lot of time. What would we be doing if we could free that time up and then finding the tools that fit that?
[37:12]
C
That's really profound, Kira. You're so right about that. Because, you know, again, I, I know I've mentioned it before, but there's that meme that I keep seeing, that cartoon where it's a group of CEOs all clamoring, we want AI and there's someone asking to do what? And they all go, we don't know, but go faster. And it's just, it is really true. It's like this, this feeling of we need to have tools now, now, now, but to do what? You know, how is that going to magically make everything better if you don't think first about what it is you're trying to accomplish. What's the outcome that you have in mind? I think this treatise is really interesting. You know, Kira, I'm glad you shared it with us. And we'll, we'll put a link to it on the podcast page. I think it's really interesting because it's not saying, I don't think it's saying what should happen. I think what it's saying is what is possible and what could happen. And so to me, that feels like a really important wake up call and a good topic of conversation around pretty much every board table of Is this the kind of future that we want to be architects of, or is there something that would be better for us? And us being all of us, not just a specific company or a specific management team or a specific board, but your employees, your communities, your customers, all of us. I mean, we're, we're right now, for better or for worse, we have a big giant global capitalist ecosystem. Is it perfect? No, not even close. Do we want to completely explode it without a clear indication of what we would replace it with? That's the right question to ask. I think we all have to be a little careful because it could happen to us if we're not putting guardrails in place.
[38:51]
D
And to your point, Jeremy says in the treatise, the good ending is possible, but only if we actually build it that way on purpose. Share ownership, keep the important pieces open, and make sure no single player gets to own the door that everyone else has to walk through. So we do have the opportunity now, right?
[39:08]
C
You could also spin out a scenario where so much of what we think of as work is automated that the connection between labor and having the means to sustain yourself get broken. And so you think differently about how people sustain themselves. You could have that conversation. I don't hear anybody having that conversation. So I just think like, this feels like a very, very slippery slope that's very steep and moving very quickly and so definitely worth having a conversation. I hope more people have conversations like the one we had with Dominique. And by the way, I can't say enough good things about Dominique. I know I probably sound like her PR agent and I'm not. I promise you there's no commercial relationship. I just think she's really smart and I think what she's doing is very laudable. And so, you know, if you haven't checked out Global Data Innovation, check them out. And if you haven't thought about doing something like the trust framework, think about it because I just think, you know, we, we all have a moment right now when we can do this better and protect ourselves, protect each other, protect our companies, protect our shareholders, protect our employees. But it won't be good if we don't do this. I think things are innovating so quickly, it's gonna spiral out of control. We're gonna be right back where we are right now with cyber, right back where we are right now with social media. We've gotten this wrong a whole lot more than we've ever gotten this right.
[40:34]
E
All right, well, let's get it right then.
[40:37]
C
Let's do it. Come on, people. Lock arms. Yay, corporate governance. All right, on that note, that wraps up another episode of the Corporate Director Podcast, the voice of modern governance. Like to say a few special thank yous. First and foremost to our amazing, trustworthy AI expert Dominique Shelton Leipzig podcast producers Kira Ciccarelli, Steve Clayden, and Terry Tirrey, our sponsors for the show, including kpmg, Wilson Sonsini, and Meridian Compensation Partners. And most especially, thank you to Diligent for sponsoring the show. If you like our show, please be sure to give us a rating on your podcast player of choice. You can also listen to our episodes and see more from Diligent Institute by going to diligent.com resources. Thank you so much for listening.
[41:19]
B
You've been listening to the Corporate Director Podcast. To ensure that you never miss an episode, subscribe to the show in your favorite podcast cast player. If you'd like to learn more about corporate governance and tools to help directors do their job better, visit www.digent.com. thank you so much for listening. Until next.

Summary7 min read

The Corporate Director Podcast

Episode: The TRUST Framework and Guardrails for AI

Date: July 29, 2026
Host: Diligent
Main Guests:

Dottie Schindlinger (Host, Executive Director, Diligent Institute)
Megan Day (Strategy Leader, Diligent)
Kira Ciccarelli (Head of Research, Diligent Institute)
Dominique Shelton Leipzig (Founder & CEO, Global Data Innovation)

Episode Overview

This episode focuses on board governance in the age of AI, examining the balance between innovation and risk and emphasizing the urgent need for effective guardrails and frameworks. The centerpiece is an in-depth interview with Dominique Shelton Leipzig, who introduces the TRUST Framework—a practical model for boards to oversee AI responsibly. The conversation explores the implications of the “agentic economy,” the rapid advances in AI-driven workflows, and the EU AI Act’s regulatory milestones. Key insights address the risks of concentrating AI power, the need for proactive governance, and concrete steps boards should take now.

Key Discussion Points & Insights

1. The Agentic Economy & AI Risks (00:16–08:03)

Emerging Agentic Economy: The hosts discuss Jeremy Allaire's "treatise on the agentic economy,” in which AI agents autonomously manage other agents and potentially transform (or destabilize) corporate structures.
- Key concern: What happens to the workforce when AI agents replace many roles, concentrating power among top decision-makers?
- Notable quote:
  “If everything is handed off to agents and you only need one person overseeing a mass of agents, then it is very likely that power will become very concentrated in the hands of very few people.” — Megan Day (04:10)
Speed of Change: AI is evolving faster than regulatory or governance structures can keep up.
- “If we don't put guardrails in place right now...then it will be an existential threat within a generation.”—Dottie Schindlinger (06:23)
Call for Action: Without immediate intervention, future leaders may lack necessary experience and preparation.

2. The Digital Trust Summit & Learning from Cybersecurity (09:16–13:17)

Summit Origins: Dominique Shelton Leipzig describes founding the Digital Trust Summit and her company to help leaders avoid past mistakes from cybersecurity, where insufficient early action led to massive risk and costs.
- “From those days in Cyber 25 years ago, I kind of vowed to myself I would never let that happen again...In the past 10 years, over $60 trillion has gone in cyber related fines and expenses…”—Dominique Shelton Leipzig (10:08)
Summit Goals: Facilitate clarity for CEOs and boards on actionable AI oversight—not just hype.
- “Create a space where CEOs, boards, regulators…could put our heads together about solutions, not hyped.”—Dominique Shelton Leipzig (11:44)

3. The EU AI Act: What Boards Need to Know (13:17–16:22)

Overview: The EU AI Act is a landmark law with sweeping scope (1000+ KPIs, 700+ pages).
- August 2nd, 2026: Major regulatory milestone date.
Focus on What Matters: Out of 1,000 KPIs, about five drive 85% of real-world AI incidents.
- “What we wanted to do with what we want directors and CEOs to focus on…are the things that…deserve my attention so that I can avoid or help our management team think through how to avoid losses that have happened to maybe 95% of the companies deploying AI…”—Dominique Shelton Leipzig (14:44)

4. Questions Boards Should Ask About AI (16:22–17:29)

Practical Board Agenda:
- Ask management to list current AI use cases.
- For each, ask: How does this use case align with our business strategy, revenue, and operations?
- Avoid “AI science projects” that don’t directly add value.
Notable Quote:
“Are we doing AI that's really going to move the needle…or are we doing science projects…we’re not going to be able to demonstrate…added value?”—Dominique Shelton Leipzig (17:19)

5. The TRUST Framework for AI Governance (17:29–21:31)

Framework Explained:
TRUST = Triage / Right Data / Uninterrupted monitoring / Supervision / Technical documentation.
- Developed by analyzing thousands of real AI incidents and legal frameworks globally.
- Focuses on areas where lack of action caused 85% of losses.
- Designed to surface board-level priorities, cutting through complex compliance requirements.
Highlights within TRUST:
- R (Right Data): Ensure data accuracy, IP rights, privacy, and business rights are covered; massive financial and brand risk if overlooked.
- U (Uninterrupted Monitoring): Continuous testing, monitoring, and auditing—can often be implemented cost-effectively.
- Notable Quote:
  “If a board or a CEO focuses their attention on innovating to anticipate and avoid those issues, then they’re to be on the road to success with their AI.” — Dominique Shelton Leipzig (19:53)

6. Making AI Guardrails Practical & Affordable (21:12–24:34)

Pushback on Costs: Boards may resist uninterrupted monitoring as expensive, but implementing guardrails is “couch cushion money” compared to AI investments and potential losses.
- “In that context, a six figure guardrail is literally a drop in the bucket.” — Dominique Shelton Leipzig (22:33)
Real-World Consequences: CEOs stepping down, directors named in lawsuits, due to lack of AI oversight.

7. AI in the Boardroom & First Steps for Directors (24:34–26:47)

Immediate Recommendations:
- Within 30 days: Adopt an AI board policy; templates are available (including from Diligent).
- Training: Ensure board members understand the “why” behind the policy through real-life stories and examples.
- Risk Example: Directors using commercial AI tools risk data leakage, even without inputting confidential data.
- Quote:
  “Risk stands…Think about this. We already know of several public companies where this has happened, where board members have used…commercially available AI…competitors have been able to derive intelligence…” — Dominique Shelton Leipzig (25:14)

8. Looking Ahead: The Future of Boardrooms (27:05–30:53)

Biggest Change in 10 Years: Only boards that embrace both innovation and strong governance will survive and thrive.
- “There’s about to be a huge schism…between those boards that proactively lean into the technology and…governance…and the others that are not going to make it” — Dominique Shelton Leipzig (27:12)
Continuous Learning: Board members must not abdicate oversight to tech experts, but stay engaged and informed.

9. Closing Reflections & Notable Moments (32:13–40:33)

Learning from the Past:
- Urgency to not repeat cyber’s mistakes—act now to prevent AI-related crises.
- “We have a chance now to get it right and to learn from previous instances where new regulations or new technologies came to pass and maybe we weren't as prepared…” — Kira Ciccarelli (32:18)
Governance as Enabler, Not Barrier:
- Good guardrails are not anti-innovation—they make innovation “move on smoother rails.”
- “Good governance is good governance when it fuels innovation…If you don’t have rails, it can just spiral out of control…”—Dottie Schindlinger (35:52)
Founders' Passion: Dominique is building next-gen governance talent through her Career Ascend program.

Memorable Quotes & Key Timestamps

Power Concentration Risk (04:10):
“Very likely that power will become very concentrated in the hands of very few people.” — Megan Day
Guardrails as Investment, Not Cost (22:33):
“In that context, a six figure guardrail is literally a drop in the bucket.” — Dominique Shelton Leipzig
Consequences of Inaction (27:12):
“There’s about to be a huge schism…between those boards that proactively lean into the technology and…governance…and the others that are not going to make it” — Dominique Shelton Leipzig
Board AI Risk Example (25:14):
“Competitors have been able to derive intelligence about those boards and about the companies that those board members are on by virtue of the prompts…” — Dominique Shelton Leipzig
Governance as Innovation Enabler (35:52):
“Good governance is good governance when it fuels innovation…If you don’t have rails, it can just spiral out of control…”—Dottie Schindlinger
Call to Action for the Future (40:33):
“We all have a moment right now when we can do this better and protect ourselves, protect each other, protect our companies…”—Dottie Schindlinger

Section Timestamps

[00:16–08:03] Opening, Agentic Economy & Existential AI Risks
[09:16–13:17] Dominique Shelton Leipzig: Motivation and Summit Origins
[13:17–16:22] EU AI Act: What Boards Must Know
[16:22–17:29] Questions Boards Should Ask Management
[17:29–21:31] The TRUST Framework Unpacked
[21:12–24:34] How to Approach Costs & Implementing Guardrails
[24:34–26:47] First Steps: Board AI Policy & Training
[27:05–30:53] Dominique’s Vision for Boardrooms of the Future
[32:13–40:33] Reflections, Governance & Innovation, Final Thoughts

Final Takeaways

Boards must balance immediate, practical AI governance with strategic oversight.
Frameworks like TRUST are crucial to distilling complex regulations into actionable board priorities.
Proactive, not reactive, guardrails will separate future-proof boards from those at risk.
Governance is not a brake on innovation—it’s the foundation that allows organizations to innovate safely and at scale.
The window to “get it right” is now; learning from the history of cybersecurity failures is critical.

Recommended Resources from the episode:

Global Data Innovation (TRUST Framework, country legal frameworks)
Diligent Institute: AI Board Policy Template
Reading: Jeremy Allaire’s treatise on the agentic economy, “What is Claude Anthropic? Doesn’t Know Either” (The New Yorker), Pope Francis’s encyclical “Humanitas Magnificus,” Haruki Murakami’s “IQ84.”

Loading summary

Transcript64 lines

[00:00]
A
Foreign.
[00:09]
B
Welcome to the Corporate Director Podcast where we discuss the experiences and ideas behind what's working in corporate board governance in our digital tech fueled world. Here you'll discover new insights from corporate leaders and governance researchers with compelling stories about corporate governance strategy, board culture, risk management, digital transformation and more.
[00:33]
C
Hi everybody and welcome back to the Corporate Director Podcast, the voice of modern governance. My name is Dottie Schindlinger, Executive director of the Diligent Institute, and I'm joined once again by my two amazing co hosts, Megan Day, strategy leader here at Diligent, and Kira Ciccarelli, who's the head of research at Diligent Institute. Megan, Kira, how are you both doing today? Very well, Dottie.
[00:55]
D
Pretty good.
[00:56]
C
So hopefully everybody had a nice weekend. I know here in the northeast of the U.S. we've just come through lots of bad air quality from the really awful, awful fires up in Ontario and then we had major storms again. Boy, this has been, this has been a summer for the record books, hasn't it? It's been crazy here.
[01:13]
D
And by the way, don't buy the produce, I guess for a little while.
[01:16]
C
I know on Saturday we had this idea of like going to a local salad bar and then three seconds later went, maybe not the time to be eating fresh lettuce. Maybe in a couple weeks. Maybe in a couple weeks. But listen, Kira, I want to give you the floor because you flagged us on an article that you saw in today's CEO Daily about Jeremy Allaire's treatise on the agentic economy. Do you want to kind of talk about that a little bit?
[01:40]
D
Yeah. So this absolutely caught my eye this morning when I was going through my newsletters and my inbox this morning. And essentially, Jeremy Allaire's tech company CEO, he and his team have put together a sort of treatise on the agentic economy. And basically the point that he is making here is that as AI agents become more dominant in workflows and just in the economy and business world in general, they're going to have to start operating with sort of blockchain and things like that, as opposed to traditional means of changing money back and forth. And it was really, really fascinating.
[02:24]
C
Yeah, I have to be honest, there's lots of different versions you can read. So if you go to the website and you check out this treatise, there's the 62nd thesis, there's the Tell it to Me like I'm a 12 year old version, there's the long read, there's the podcast, there's a video. I went to the Tell it to Me, like I'm a 12 year old version even that I found really challenging. But the thing that I was struck with. So he's talking about agents managing agents and you know, dividing jobs into discrete tasks that they can assign to each other. And the way that it's written made me think, so is this the end of corporations as we know it, part one and part two? Where in this treatise does he cover what happens to the rest of us? Because he talks about how, you know, agentic AI can't replace hands on work. So people who, you know are caregivers or, you know, sanitation workers or anybody who's working with their hands, their jobs are probably safe for now because robotics is still a long way away from replacing them. And all the people at the top are okay because they, you know, you still have to have a human being taking accountability, responsibility. But the rest of us, question mark, did he cover it anywhere in the street?
[03:41]
D
So I like you, I started with the 62nd version and then I was thinking to myself, none of these words are making any sense to me at all. So I quickly went over to the. Explain it like I'm 12. And yeah, I think you're right. I mean, he makes the point in this five minute version that's kind of covering some of the major tenets of the treatise, that the huge risk here is exactly what you're saying. If everything is handed off to agents and you only need one person overseeing a mass of agents, then it is very likely that power will become very concentrated in the hands of very few people.
[04:27]
C
So that doesn't sound like a great idea to me. I, you know, it's, it's really interesting as I read through it, as I thought about it some more, I mean, in fairness, Jeremy Allaire, CEO of a company that is in the blockchain space. So maybe not super surprising that he's anticipating that blockchain will be part of the ecosystem that makes this all possible and happen. But there were certain phrases that really kind of set me back, like this will happen and it'll seem like it was almost by accident. There was a phrase in there that said that, and I thought, you know, that is really important for us to pause on and say that will be true if we don't do anything right now. Megan, I don't know what your take on all this is, but yeah, what
[05:13]
E
I struggle with, and this has come up in some of our other conversations is how soon is this reality going to be here? And, and I oscillate between, oh my God, the End is near to. All right, we have some time here. This is, you know, 10, 15, 20 years out before we live in this society. I do encourage folks, if you haven't read it, earlier this year in the New Yorker, there was a great piece called what is Claude Anthropic? Doesn't know either. And it's worth reading on a couple of fronts. One, it sends you into the mind, if you will, of this agent economy and some of the things that we might grapple with in the future. But like the manifesto, the treatise, it does pose a million different questions about where we're headed and how quickly we will get there.
[06:11]
C
I mean, one of the things I think about, and this is probably a really good segue for us to listen to our interview for today, but one of the things I think about is, you know, we're seeing signs that this is already starting. And so to your question, Megan, about how soon is it before we're in this new world? Well, we will get there. If we don't put guardrails in place right now. Like, if we don't, if we don't, like, have some conversations at the right leadership level to talk about how do we keep this from being an existential threat, then it will be an existential threat within a generation. Like, within a generation. If what this treatise lays out comes to pass and nothing else changes, that's the most important thing. If nothing else changes within a generation, you have absolutely no pipeline for future leadership. Within a generation, you have absolutely no one who is prepared and qualified to be those people at the top running the agent economy.
[07:03]
D
Right.
[07:04]
C
Truthfully. Because how do we, how do we prepare people today? They come up through the ranks, they learn entry level jobs, then they get more responsibility, then they manage some people, then they get more responsibility, and over the course of years, they accrete enough experience and an education that they are now ready to lead teams, departments, companies, that whole pathway. If we were to just allow the agentic economy to happen the way it's written about in the treatise completely breaks down and whole companies, whole types of jobs, whole sectors of the economy all become agentic with robots governing other robots. What are the people doing? I'm seeing you both nod your heads and saying absolutely nothing, which is terrifying.
[07:49]
E
To even remind me of the other manifesto, if you will, from earlier this year. And I'm blanking on who wrote it, but we talked about it in the podcast.
[07:57]
A
That's it.
[07:58]
E
I mean, it's terrifying as it is exciting.
[08:03]
C
So that's probably a very good segue to A Glimmer of Hope, which is the event that Kira and I got a chance to go to in June in Brussels. We were invited to come and speak at the Digital Trust Summit that was organized by Dominique Shelton Leipzig. And so today we sat down with Dominique, we asked her lots of questions about the event, about why it's important, about why every director should think about trying to go next year, every CEO should think about trying to go next year. And this treatise I think just adds extra fuel to that fire for me that we really all, we will all do ourselves a disservice if we don't get this right. And pretty much right now we need to get the right guardrails in place. And the great thing about this event was it was all about guardrails and innovation. It wasn't about how guardrails slowed down innovation, it was about how guardrails speed innovation. And I think that's really cool. So why don't we give a listen to the interview and we can come back and maybe the two of you will have more to say. Joining us on the Corporate Director podcast today is Dominique Shelton Leipzig, founder and CEO of Global Data Innovation. Dominique, thank you so much for joining us again on the show.
[09:17]
A
Dottie, thank you so much for having me. It's great to be back.
[09:20]
C
Well, let's just get started off here by having you introduce a bit more about yourself to our listeners than your current title. You've had quite an amazing career, so could you share a little bit about your career background?
[09:32]
A
Yeah. So you know, I started my company, Global Data Innovation almost three years ago now because I could see that we were getting with AI and new technologies kind of in the same pickle we had gotten into with cyber. In other words, there is new technology out there, confusion at the board level and CEO level about what to, what to do with this new technology, how to govern it. And then what I noticed was a tendency to over delegate to anyone else that would take this on to figure it out. And what that leaves is a huge bucket of risk. So from those days in Cyber 25 years ago, I kind of vowed to myself I would never let that happen again. Because of course, cyber has turned into an issue that has cost our global economy about 12 and a half trillion per year. And in the past 10 years, over 60 trillion has gone in cyber related fines and expenses and ransomware payments, et cetera. And when I saw those trends starting to happen with AI about three years ago, I realized that we needed to do something we didn't do with cyber, which is leverage best practices solutions and really our study of AI failures, bringing those learnings to our leaders ahead of time so they could map solutions to avoid those problems and be successful with their technology.
[11:05]
C
Well, I know that the work that you're doing is really having a big impact because myself and Kira Ciccarelli, my colleague, had the great privilege to join you in Brussels just a few weeks ago at the Digital Trust Summit. And I think this is the fourth year, is that correct? Maybe the fourth or the fifth year?
[11:23]
A
The fourth, yeah.
[11:25]
C
So we were so excited to be there. Could you share with our audience a little bit more about the goals of the summit, kind of what led you to start this event in the first place and what were some of the biggest themes or wake up calls that were shared at this year's event?
[11:39]
A
Yeah, so Donny, I started the Digital Trust summit in 2023, very close, about a year before I started my company. And I started it for the very same reason I started my company. I could see, see that our CEO and board community were a little bit deers in headlights with the new AI technology and through no fault of their own. The technology is moving so fast, the trends are moving so quickly and there were so many headlines even back then. It was clear to me that we needed to create a space where our CEOs boards could get together with regulators, legislators and large language model experts, data scientists and academia, and put our heads together about solutions not hyped. So that mission that started in 2023 continues and remains as you know, the Dottie, because you do so much of this work at the Diligent Institute, we are literally our, our CEO and board members are inundated with all kinds of signals, headlines, noise. It's very hard to make, make sense of it all and to figure out what should boards be doing right now to govern AI. So the. The goal of this conference was really to provide in clarity and a roadmap for solutions for excellent exercise of fiduciary duty over AI. And I believe we accomplished that. So I'm very proud that you were there and you were a big part of that. Dottie, in terms of your panel on laying out what boards in the EU and US need to know about trends in order to arrive at great governance.
[13:17]
C
Well, speaking of boards in the EU and the us we are just days away from a big milestone in the EU AI Act. I know August 2nd represent a really important day, so could you share a little bit about what directors need to know most about the EU AI act and what's happening on on August 2nd that they need to be aware of?
[13:38]
A
Yes. So for directors that are directors of companies that either do business in the eu, have vendors in the eu, or see the EU as a growth market, which is a big swath of our US companies, if you're on a board of a company like that, the EU AI act should be on your radar screen on August 2nd. First of all, the EU AI act is a sweeping law. There's over 1,000 KPIs in it, things that are very hefty, about 700 pages or so of law in there, a lot going on. But what we've done at Global Data Innovation is taken that law and looked at what aspects of this law are most germane to avoid the tens of thousands of incidents that we've been studying. As I mentioned, there are so many incidents that have occurred with AI in the four years since ChatGPT hit the market. And from those learnings, we've seen that out of those thousand KPIs in the EU AI act, there are only about five that really are moving the needle and causing about 85% of the incidents that have been documented. The tens of thousands of incidents out there. So what we wanted to do with what we want directors and CEOs to focus on at the level that they're in, obviously their legal counsel will make sure that they're buttoned up and all the legal compliance. But what a CEO and a board member needs to know, what are the things that are going to, that deserve my attention so that I can avoid or help our management team think through how to avoid losses that have happened to maybe 95% of the companies deploying AI, so that we do not repeat those incidents in our company? So we want incidents to stop where they are, not repeat themselves over and over again in companies in various ways, like what we have with cyber.
[15:42]
C
So, you know, a couple of things that you just touched on there are really resonating for me because I think what always happens when we have a new technology come to play, the development of that technology and the adoption of the technology almost always outstrips the governance of the technology. And I think we're absolutely seeing that with AI, that AI is moving a lot faster than regulation is able to keep up, and really, even organizational governance is keeping up. So what are some of the questions that boards should be asking management right now to really get a sense of where AI is being used across their businesses, including shadow or embedded AI? What, what should they be asking about yeah.
[16:23]
A
You know, first, every board member should be asking for management to make sure that AI is on the agenda in a specific way. To, to your point, do what AI use cases do we have going on in the business? That is a great question to get a report on. And then similarly, once that report is delivered, having the resilience or pushback to ask, what do these AI use cases that are in flight? What do they have to do with our strategy in the company in terms of revenue, operations and overhead cost reductions? What impact will this AI have on our business if it is successful? So getting aligned on those preliminary questions really helps the board get a sense of are we doing AI that's really going to move the needle if it's, if it's successful, or are we doing science projects in the company that even if we have home runs with it, we're not going to be able to demonstrate to an investor or to our stakeholders, our business partners and employees and customers that this is added value.
[17:29]
C
So, Dominique, one of the things I was eager to talk to you about is the trust framework that you developed. And trust T rust is an acronym that stands for triage.
[17:39]
D
Right.
[17:40]
C
Data, uninterrupted monitoring, supervision and technical documentation. And first of all, I love that framework so much because it's so easy to understand and get your head around. But I wanted to ask you more specifically about the genesis of that framework and how that maps to the EU AI Act. And also just where do you see companies falling short on that framework today?
[18:06]
A
Yeah, so the genesis of the framework was really ground up. I looked at AI incidents that were occurring and then compared them with where I saw commonalities in the legal frameworks in 100 countries, six continents. A lot of those frameworks are on my company's website@globaldatainnovation.com there's a frameworks tab and you can see all the countries that have legal frameworks. And what I noticed is there were certain themes that kept coming up in all of the legislation. I also then we took the four years now since almost four years of data we have since ChatGPT4 was first announced. And there are tens of thousands of incidents. So what we did is we mapped where we saw commonality in the law with what would have prevented those tens of thousands of incidents from occurring had those best practices, some of the laws aren't in. In effect yet. Had those best practices been enacted, what would have prevented those losses? And it came down to these five areas. You know, when I think of the EU AI act, which has over a thousand KPIs or NIST, which many companies are mapping to just The NIST Gen AI framework alone has over 260 KPIs. That's so much for a board or a CEO or even your general counsel to find focus on all at once. So I wanted to find a clean way to prioritize what is really important. And what we learned is, is that about 85% of the incidents are caused by the lack of something in the T area, the R area, the U area, the S area, the T area that you outlined, Dottie. So that's why we focus on it, because we figure that if a board or a CEO focuses their attention on innovating to anticipate and avoid those issues, then they' to be on the road to success with their AI.
[20:07]
C
So I love that so much. You know, of course if you're on a board, you're going to want to ask what kind of framework is in place around governance of AI. But then are there a couple of the letters that board members should be asking about at every meeting? I mean, are there specific things they should be getting reports on at every meeting?
[20:25]
A
Yeah, I think the two most important and where we've seen the biggest losses are in the area of R, the right data to train. And that's really about data accuracy, making sure that there's a process for ensuring that accurate data is used for the AI. And then second in the right data has to do with IP rights, privacy rights and business rights to train. We're seeing billions of dollars going into settlements on for copyright and other IP related claims. And so it's an area of focus that we think boards should pay attention to. Another area for that the board should pay attention to here is, is and I call it like the fail safe is the untrust, uninterrupted testing, monitoring and auditing.
[21:12]
C
And I would imagine part of the reason why boards should be asking about those two areas is that especially in terms of the uninterrupted monitoring, that that might sound like something that is costly and so there might be pushback on it. But what would, what would you counsel a company that tells you such a thing?
[21:31]
A
Well, the good news is times have changed. AI is here and uninterrupted testing, monitoring and auditing. I've seen certain guardrails, you know, say there's monitoring against a known accuracy standard that the company has for say if it's a genomic company for certain types of R and D for disease, or if it's a retailer, certain standards they have for customer service. The good news is that with AI, these standards can be coded into the AI itself without using generative AI. They can be used using deterministic or what I mean by deterministic, just hard coded rules that don't need to shift with the Internet or the waves of the whole wide Internet. These are your company's standards with respect to accuracy. So they can live in uninterrupted testing, monitoring and auditing in a hard coded way. They can live either in what something called an observability model, or they can live in the AI itself as a monitoring vehicle. And we've seen that done as low as five figures, six figures. So let me put this in context. If you're working for, you're on the board of a big bank or a healthcare company, they are putting in, say it's a small mid cap healthcare company, they're putting in millions of dollars associated with AI development and some of our bigger banks anywhere between, you know, the proxies reveal and their annual reports, anywhere between, you know, 15 billion, 19 billion in digital transformation and AI development. In that context, a six figure guardrail is literally a drop in the bucket. And when I think of the CEOs just this year that have had to step down because of AI incidents and the shareholder derivative actions that are naming board members individually for failing to exercise oversight and the months that those directors will be sitting in deposition and looking through their emails and sitting with counsel, it is so worth it to take care of these issues and ask the question as a board member now rather than be defending your governance and your oversight later.
[23:50]
C
Such good advice. Well, one of the other aspects of AI and governance is the use of AI in board work itself, which of course is something that we spoke about at the Digital Trust Summit. Diligent Institute's been doing a lot of research in this area and one of the things that we have found is that while directors are using AI pretty, pretty frequently to help them with board work, very few companies have AI usage policies in place that governor usage of AI specifically in the boardroom. So if you were advising a board that you know, has been experimenting a little bit with AI, but still doesn't have a very mature governance framework, what would be some of the first things that you'd recommend directors do in the Next, let's say 60 to 90 days to really get ahead of the risk?
[24:34]
A
Well first, within the first 30 days, I would highly recommend that the board get to the Diligent portal and pull down the sample template AI Board policy that you've made available, Dottie, which is incredible, but it's absolutely important for the board to have their own AI governance policy. Why? Because in the interstitial area of boards using AI and the lack of guidance on what it should be used for and should not be used for, risk stands. You know, think about this. We already know of several public companies where this has happened, where they've board members have used AI commercially available AI, not enterprise licenses, and competitors have been able to derive intelligence about those boards and about the companies that those board members are on by virtue of the prompts, even if they don't even put in confidential information by virtue of the prompts that they use. So this is so important for boards to have that guidance and to focus on making sure that there's a policy in place. Second step is training on that policy, so everybody knows why it's there. Dottie, you and I have done a number of trainings together and I, I. One of the things I like doing with you so much is the fact that both of us bring up a lot of stories. I find it so much easier for a board member to remember the importance of why it's important to use an approved portal that has security controls like what you've got going on at Diligent versus the general commercial tools that are available and they're not necessarily confidential. And the providers themselves say, please don't put anything in here or use prompts that you would be concerned about being the subject of a data breach, because many of the providers have suffered data breaches. That's the reason they tell you that in their terms. So I love to do that storytelling. So our board members have the, not just the, the policy itself, but the why it's important and what it's there to protect.
[26:47]
C
Such, such good advice. Well, listen, before we wrap up our conversation here, there are three questions that we'd like to ask every guest that joins us on the show. And I know you've answered these before, but we're going to ask them again. So the first question is, what do you think will be the biggest difference between boardrooms today and 10 years from now?
[27:05]
A
Well, I think you're going to see the boards that are around 10 years from now are going to be the ones that adopt these proactive practices that I'm talking about here. There's about to be a huge schism that lays out in the next, I would say 12 months to 24 months between those boards that proactively lean into the technology and also equally Important lean into the governance so that they're turning out AI products that have a level of, that match the brand, that have a level of excellence, accuracy and really reflect what's best about your brand. There'll be the others that are not going to make it because they're going to turn out products and AI enabled material that's inaccurate, that causes brand tarnishment and harm to their customers, employees and business partners. And they're not going to be sitting with us ten years from now. So I'm excited that we're going to have a really sophisticated group of companies and boards that have learned how to manage this technology and many more new things. We're just at the beginning of Quantum and we're at the nascents of what AI will be doing in agents and agentic. AI and robotics saw so many prototypes at the UN and Geneva of all the new technology that's coming our way. But the boards that become flexible in this reflex of being continuous learners, asking questions, not deferring and putting our heads in the sand and over delegating to the technical folks, but jumping in with strategy and making sure that you understand as a board member what you're overseeing. Actually ask for the types of briefings that will help you understand it. That's the. Those are the boards of tomorrow and I'm excited about that.
[28:51]
C
Well, speaking of briefings, what was the last thing that you read or watched or listened to that made you think about governance in a new light?
[28:58]
A
There are two things. One is the Pope's encyclical letter Humanitas Magnificus, which came out in June. And I just. It's about 100 pages long. It's a long read and when you read it, it's. You can see some of the difference between AI writing and a real person's writing. Because it's clear that somebody working with the Pope and the Pope himself spent a lot of time in putting together a great document that I think eventually history will show as a roadmap. It's not really a religious document. It's about planning and thinking ahead of the society that we want with AI and how to get us from point A to point B. It talks about the dignity of work and also the importance of accuracy and truth. These are all things that are going to be essential for our future. And I'm excited that there's roadmap there in that document. So I found that that was very synergistic with the work that we do at Global Data Innovation and also an extension of which is our Digital Trust Summit where we act out of simulations and work on forward planning and thoughtful pre planning around AI. The other book I came back to is Haruki Murakami's IQ84. I highly recommend it. It's really about technology, lives, science fiction. It also has kind of a Proustian way about it. It's another long read. That one's about 1200 pages. But I read it about 10 years ago and I came back to it and I've been rereading it this summer and it's just been a joy.
[30:38]
C
Those are great, great suggestions. I think you get the prize for being the first guest to ever recommend a papal encyclical. So thank you for that. That's really fantastic. And Dominique, before we let you go, what is your current passion project?
[30:53]
A
Yeah, you know, right now I'm really passionate about making sure that we build also the next generation of governance experts. So I started a Career Ascend program as part of Global Data Innovation. We just came off our first cohort of 11amazing governance professionals and really refined many of the things that we've been talking about here. How to use their voice, how to see something, say something, and if there's something that they notice at the ground level, to be sure to raise those things to the CEO and board so that everybody's working off the same sheet of music and we're getting to great governance, not waiting for it to be turned into something like what happened with cyber, where we kept things in silos and it ended up costing our global economy trillion. So I'm excited about Career Ascend.
[31:47]
C
Well, Dominique, I'm excited that I got a chance to speak with you again. Thank you so much for coming back and joining us on the show.
[31:52]
A
Thank you for having me. Dottie, it's been an absolute pleasure and congratulations on all the work you're doing at Diligent.
[31:58]
C
We've been joined once again by Dominique Shelton Leipzig, founder and CEO of Global Data Innovation. Dominique, thank you so much for joining us.
[32:06]
E
I was going to say stop talking. Yes, this is all good, Kira. I feel like you got it.
[32:14]
D
Great interview, Dottie. It's always such a pleasure to get to talk to Dominique and just kind of mine her brain. And I think one thing that really stuck out to me from this interview in particular is that she comes back often to this idea of we have a chance now to get it right and to learn from previous instances where new regulations or new technologies came to pass and maybe we weren't as prepared and thinking about this. Treatise on the agentic economy that we've been discussing. My first thought reading it was when you're talking about agents, AI agents operating within a capitalistic society where it's money changing hands and you have agents doing that work. I just imagine every cfo, every legal leader, just like, I can almost hear them screaming as they read it, like, from where I am. And I think that's just such a great example of where something like the trust framework, which Dominique talks about could be so, so important. And the idea of knowing the art of the possible and what might be coming and be able to set up those guardrails in advance. We're at the point now where we have the chance to do that.
[33:27]
C
These are times when this is totally going to date me, but these are times when I hear Jeff Goldblum from Jurassic park in my head saying, you were so focused on whether or not you could, you didn't pause to ask if you should kind of feel like that applies. Megan, what was your big takeaway from this interview?
[33:46]
E
Well, I will. I mean, I come back to what we talked a little bit before about governance not being anti innovation, that it can, you know, if you have weak governance, it creates paralysis and you can't scale AI enterprise wide. But she gives some really practical advice too, for boards about how to think about this, which I think again, when we talk about all of this sort of nebulous stuff, you really need to come back to, well, what can I do right now? And that's about asking at every meeting whether the company has uninterrupted testing, monitoring and auditing of these AI outputs. I mean, I think that is just, again, what we need to keep coming back to.
[34:33]
C
And I really loved her point that, you know, you might get an answer of, well, that's too expensive and cumbersome and we can't possibly do it. And she, she brought up the case of, you know, I've seen it happen for as little as five figures or six figures. She's like, you know, if you're making these multi billion dollar bets, that's like, that's couch cushion money, you know, to, to invest in having something that will keep you from an absolute existential threat. I think the other thing I always appreciate about Dominique's take on how we can get this right now, you know, she lived through her career spans across the entirety of the cybersecurity world and she saw how there was a moment in the 90s when we could have gotten ahead of it and had we put some frameworks in place at that point in time, we wouldn't have CyberCrime as a $12 trillion a year industry right now. And I think nobody wants a repeat of that. Like literally no one wants something like that to happen with AI. Nobody. That's not good for anyone except maybe the criminals. It's not good for anyone else. And it has caused nothing but headaches for companies, for consumers, for employees, for everyone. And I think that's such a good analogy. And I just keep holding onto that because I think that's what's in it for us.
[35:45]
D
Right.
[35:46]
C
You know, to your point, I don't think good governance slows down innovation. I personally think good governance is good governance when it fuels innovation, when it allows innovation to move on smoother Rails. Because you have rails. Like if you don't have Rails, it can just spiral out of control and pretty soon you end up with a company that is ending because it's done something really terrible without your intention, and now it's terrible. And now you have to clean it up versus good rails allow you to go smooth and fast in a direction.
[36:18]
E
And granted, I am not in a public company boardroom at the moment, but I have to think a lot of the conversations are focused on perhaps you need to be using AI more. Add it to this use case. Add it to that use case around the business versus coming at it with a skeptics lens of how does this connect to strategy, how does this connect to operations? How is this creating value for the organization outside of maybe cutting heads? So lots to think about.
[36:50]
D
I hear that all the time. And I think we see a lot of organizations move in the wrong order. Like they start with, we have this tool now, so where can you force fit this in? Where can you plug this in? Instead of starting with what are we doing? That is tedious and takes a lot of time. What would we be doing if we could free that time up and then finding the tools that fit that?
[37:12]
C
That's really profound, Kira. You're so right about that. Because, you know, again, I, I know I've mentioned it before, but there's that meme that I keep seeing, that cartoon where it's a group of CEOs all clamoring, we want AI and there's someone asking to do what? And they all go, we don't know, but go faster. And it's just, it is really true. It's like this, this feeling of we need to have tools now, now, now, but to do what? You know, how is that going to magically make everything better if you don't think first about what it is you're trying to accomplish. What's the outcome that you have in mind? I think this treatise is really interesting. You know, Kira, I'm glad you shared it with us. And we'll, we'll put a link to it on the podcast page. I think it's really interesting because it's not saying, I don't think it's saying what should happen. I think what it's saying is what is possible and what could happen. And so to me, that feels like a really important wake up call and a good topic of conversation around pretty much every board table of Is this the kind of future that we want to be architects of, or is there something that would be better for us? And us being all of us, not just a specific company or a specific management team or a specific board, but your employees, your communities, your customers, all of us. I mean, we're, we're right now, for better or for worse, we have a big giant global capitalist ecosystem. Is it perfect? No, not even close. Do we want to completely explode it without a clear indication of what we would replace it with? That's the right question to ask. I think we all have to be a little careful because it could happen to us if we're not putting guardrails in place.
[38:51]
D
And to your point, Jeremy says in the treatise, the good ending is possible, but only if we actually build it that way on purpose. Share ownership, keep the important pieces open, and make sure no single player gets to own the door that everyone else has to walk through. So we do have the opportunity now, right?
[39:08]
C
You could also spin out a scenario where so much of what we think of as work is automated that the connection between labor and having the means to sustain yourself get broken. And so you think differently about how people sustain themselves. You could have that conversation. I don't hear anybody having that conversation. So I just think like, this feels like a very, very slippery slope that's very steep and moving very quickly and so definitely worth having a conversation. I hope more people have conversations like the one we had with Dominique. And by the way, I can't say enough good things about Dominique. I know I probably sound like her PR agent and I'm not. I promise you there's no commercial relationship. I just think she's really smart and I think what she's doing is very laudable. And so, you know, if you haven't checked out Global Data Innovation, check them out. And if you haven't thought about doing something like the trust framework, think about it because I just think, you know, we, we all have a moment right now when we can do this better and protect ourselves, protect each other, protect our companies, protect our shareholders, protect our employees. But it won't be good if we don't do this. I think things are innovating so quickly, it's gonna spiral out of control. We're gonna be right back where we are right now with cyber, right back where we are right now with social media. We've gotten this wrong a whole lot more than we've ever gotten this right.
[40:34]
E
All right, well, let's get it right then.
[40:37]
C
Let's do it. Come on, people. Lock arms. Yay, corporate governance. All right, on that note, that wraps up another episode of the Corporate Director Podcast, the voice of modern governance. Like to say a few special thank yous. First and foremost to our amazing, trustworthy AI expert Dominique Shelton Leipzig podcast producers Kira Ciccarelli, Steve Clayden, and Terry Tirrey, our sponsors for the show, including kpmg, Wilson Sonsini, and Meridian Compensation Partners. And most especially, thank you to Diligent for sponsoring the show. If you like our show, please be sure to give us a rating on your podcast player of choice. You can also listen to our episodes and see more from Diligent Institute by going to diligent.com resources. Thank you so much for listening.
[41:19]
B
You've been listening to the Corporate Director Podcast. To ensure that you never miss an episode, subscribe to the show in your favorite podcast cast player. If you'd like to learn more about corporate governance and tools to help directors do their job better, visit www.digent.com. thank you so much for listening. Until next.

The TRUST framework and guardrails for AI

The TRUST framework and guardrails for AI

The Corporate Director Podcast

Episode: The TRUST Framework and Guardrails for AI

Episode Overview

Key Discussion Points & Insights

1. The Agentic Economy & AI Risks (00:16–08:03)

2. The Digital Trust Summit & Learning from Cybersecurity (09:16–13:17)

3. The EU AI Act: What Boards Need to Know (13:17–16:22)

4. Questions Boards Should Ask About AI (16:22–17:29)

5. The TRUST Framework for AI Governance (17:29–21:31)

6. Making AI Guardrails Practical & Affordable (21:12–24:34)

7. AI in the Boardroom & First Steps for Directors (24:34–26:47)

8. Looking Ahead: The Future of Boardrooms (27:05–30:53)

9. Closing Reflections & Notable Moments (32:13–40:33)

Memorable Quotes & Key Timestamps

Section Timestamps

Final Takeaways

The Corporate Director Podcast

Episode: The TRUST Framework and Guardrails for AI

Episode Overview

Key Discussion Points & Insights

1. The Agentic Economy & AI Risks (00:16–08:03)

2. The Digital Trust Summit & Learning from Cybersecurity (09:16–13:17)

3. The EU AI Act: What Boards Need to Know (13:17–16:22)

4. Questions Boards Should Ask About AI (16:22–17:29)

5. The TRUST Framework for AI Governance (17:29–21:31)

6. Making AI Guardrails Practical & Affordable (21:12–24:34)

7. AI in the Boardroom & First Steps for Directors (24:34–26:47)

8. Looking Ahead: The Future of Boardrooms (27:05–30:53)

9. Closing Reflections & Notable Moments (32:13–40:33)

Memorable Quotes & Key Timestamps

Section Timestamps

Final Takeaways